Extractacore 90% less waste in vegetable processing
The Challenge
Hand-cut de-coring is inherently inconsistent. Every head of lettuce, cabbage, or broccoli varies in size and shape, and operator fatigue compounds that variation across a shift. The result across manual lines: 75 to 85% usable yield, meaning up to a quarter of good product is lost at a single processing step.
We integrated Robovision's Indurion Suite into our STAQ-software (See-Think-Act), and over the years it evolved with us, continuing to provide the flexibility we need.
The solution
Qing built Extractacore's de-coring cells by combining their own proprietary STAQ software with Indurion Suite, giving food-grade robotics the industrial vision intelligence to track each head individually rather than apply one fixed cut. The cell reads the actual size and shape in front of it and removes only the core, adapting head to head instead of forcing every product through the same pattern.
It runs in food-grade 304L/316L stainless steel, built for wash-down environments, and drops into existing lines in single or twin-lane configurations.
Strategic advantages of an integrated AI platform for an OEM
Food machine builders face a hard choice when a customer demands better accuracy: bolt on another point solution, or rebuild the vision and control stack from scratch. Neither works well for an OEM that has already invested years into its mechanical design and robotics.
An integrated AI platform solves a different problem. It brings vision, decision-making, and robotic control under one system, so an OEM does not have to stitch together separate vendors for cameras, software, and motion control. That mattered for Extractacore, whose iceberg lettuce de-coring machines had hit the limits of color-based vision alone. Extractacore could upgrade the intelligence of its machines without replacing the robot arm it had already engineered and validated. The platform approach also scales.
Once the AI layer proves itself on one machine, an OEM can extend it to other product lines without starting the integration work over again.
Technological partnership with Robovision
Extractacore did not build this capability alone. QING, a vision integration specialist, used Robovision's Indurion Suite and combined it with their software called STAQ, a control system that combines the "See" and "Think" stages of vision AI with robotic execution.
QING added a 3D camera to Extractacore's existing line and kept the robot Extractacore already relied on. The Robovision platform handles the harder problem underneath: reading the exact position and orientation of each head of lettuce, then telling the robot precisely where to cut. That division of labor let QING focus on tuning the software to Extractacore's specific process, while Robovision's Indurion Suite supplied the underlying AI and vision capability already proven across other food processing lines.
Results after implementation
The upgrade cut product waste from a range of 30 to 40 percent down to 3 to 5 percent, a drop of roughly 90 percent.
The system processes one head of lettuce per second, matching the pace Extractacore's customers need on a production floor. Because the vision AI now reads each vegetable's position and orientation before the robot moves, the line runs with more consistency across a full 24-hour shift, not just under ideal lighting or product conditions.
The deployment also generates performance data Extractacore did not have before, data it can now use to refine the process further or apply the same approach to other machines in its lineup.
Word of this success is spreading like an expanding oil slick.
Teun Keusters, Product Manager Qing Food Automation